How long should you run a Facebook ad before killing it

Most operators kill ads too early. Competitor runtime data tells you when to kill, when to scale, and when to leave an ad alone.

The 7-14 day rule is generic advice that costs you money

Open any "how to run Facebook ads" article and you get the same line. Test for 7 to 14 days.

Kill if CPA is too high. Move on.

That rule was not built for your account. It was built to be safe for a beginner reading their first blog post.

The problem is simple. Facebook's auction takes time to find your buyers.

Day 3 spend is noise. Day 7 is still noise for anything above $25 AOV.

By day 14 you finally have a signal. But it is the early signal.

The one that looks flat right before the optimization actually kicks in.

Kill at day 7 and you kill ads your competitors are happily scaling to $10k/day two months later. That is not a theory.

You can go check their runtimes yourself.

What 60, 90, and 120-day ads actually tell you

Here is the lens I use. Every active ad in the market has a shelf life, and the shelf life is a tell.

0 to 14 days. Testing.

You cannot learn anything from a competitor ad in this window. It might be about to die.

It might be their next winner.

30+ days. They have not killed it yet.

That means CPA is at worst tolerable. Interesting, not proof.

60+ days. Profitable.

Nobody in DTC funds a losing ad for two months. Media buyers get bored, budgets get pulled, CFOs ask questions.

If the ad is still running at day 60, it is making the unit economics work.

90+ days. Scaling.

The brand is actively spending on this creative and has decided it is worth defending against creative fatigue. This is usually a top-3 ad in their account.

120+ days. Category winner.

You are looking at a creative that has outlived every test the brand has run against it. Those are rare.

When you find one in your niche, you study every frame.

Short version. You do not need a formula.

You need to see which ads in your exact category have crossed 60 days, and which ones have crossed 120. Then you work backwards.

How I find the long runners in any niche

This is the move. I open Brandsearch Discovery and filter to the runtime tier I care about before I look at anything else.

The filter is called Running Days. Set it to 25+ for a broad sweep. 60+ when I want to see only profitable creatives. 90+ when I want to study scalers.

I add platform (Meta), format (Video), and the niche keyword on top of that. The grid that comes back is not a list of "cool ads you could run".

It is a list of ads your competitors have paid real money to keep alive past the point where a losing ad gets killed.

There is a difference, and it changes how you read every ad in the grid.

I sort by Total Reach next. Long runtime plus high reach is the strongest proof signal in the app.

You are filtering for ads that both the algorithm and the CFO agreed were worth keeping.

Then I click into each ad and read the hooks. Not to copy them.

To understand what angle holds attention for 60+ days in my category. Hooks that survive that long are usually doing one thing well, either a specific pain point most brands in the niche ignore, or a demo format the algorithm keeps rewarding.

Do not trust runtime alone, cross-check the brand

A single ad running 90 days can be a vanity campaign. The real signal is a brand running multiple long runners at once.

When I find an ad I want to learn from, I open the brand in Brand Analysis. The Overview tab shows ad count, traffic trend, and revenue estimate in one view.

That is the 15-second check.

If the brand has 30+ active ads, growing traffic, and a handful of creatives past 60 days, the long-running ad I found is part of a real machine. If the brand is tiny and the ad is their only long runner, treat it as one data point and move on.

This is the two-step check I run before I let a competitor's runtime shape my own decisions. Long runtime on a weak brand is anecdote.

Long runtime on a growing brand is the market telling you what works.

A real scenario

Say you are testing a supplement angle and your ads look flat at day 6. CPA is $42 and your target is $35.

Do you kill?

Under the 7-day rule, yes. You shift budget to the next test and move on.

Under the competitor runtime rule, you do a five-minute check first. Open Discovery, filter to Meta video with Running Days 60+, search your niche keyword.

Look at the shelf.

If you see three or four brands running the same angle past day 60, the angle has a ceiling well above $35 CPA. Your day 6 number is early, not broken.

You extend the test to day 14, let spend reach $800 to $1,000, and judge on that.

If nobody in your niche has a 60+ day ad on that angle, the story is different. The angle may be untested in the market, or it may have been tried and killed by better-funded accounts.

You go find out which before you put another dollar in.

Same data from Facebook. Different decision.

The tiebreaker is what the rest of the category is doing.

The reason this works is boring. Facebook's auction punishes short tests.

Short tests hit new audiences, high CPMs, cold data. Your day 6 CPA is mostly a reflection of auction friction, not a verdict on the creative.

Long-running competitor ads already paid that tax. They are living in the post-friction zone of the auction, where CPMs are lower and the audience is dialed in.

When you see an ad past day 60 in your niche, you are seeing a creative that made it through exactly the phase you are still stuck in.

That is the entire trick. You stop asking "is this ad working yet" and start asking "has anyone made this angle work at all".

The first question has no answer at day 6. The second has an answer in five minutes.

Build your own kill and scale rules from what you see

Once you have looked at a few dozen 60+ day competitor ads in your niche, your own kill thresholds get easier.

Here is the rule I use on my own account. If a competitor ad close to my angle is still running at day 60, I give my version 14 to 21 days of real spend before I judge it.

Not 7.

If a competitor has a 120+ day ad with the exact angle I am testing, I do not need a CPA target. I need to understand why that ad has not died, and whether my creative is giving Facebook enough signal to work with.

And if the shelf is empty at 60 days for the angle I am testing, I cap my test spend hard. No point running a 21-day experiment on something nobody in the market has made work yet.

The short version:

1. Pull the competitor shelf in Brandsearch Discovery, filter Running Days to 60+, sort by Total Reach.
2. Validate each long runner in Brandsearch Brand Analysis, Overview tab, check ad count and traffic trend.
3. Set your own kill window based on whether any competitor in your niche has a 60+ day ad on that angle.
4. Re-check the shelf every two weeks. The set changes. That is the point.

Summary

Stop killing ads at day 7 because a blog post said so. The market already ran the test for you.

The only question is whether you look.

Competitor ads running 60 days are profitable. 90 days are scaling. 120 days are the ones you study frame by frame.

You find all three in Brandsearch Discovery by setting Running Days and sorting by reach. You cross-check each one in Brandsearch Brand Analysis before you trust it.

Then you set your own thresholds against what is actually surviving in your category, not against a generic testing window.

The ad kill decision is not a gut call. It is a competitor lookup.